neurodudes

There’s a nice editorial in Nature Neuroscience about the Broad Institute’s PsychHTS initiative. The initiative invites scientists from outside the Broad to suggest new high-throughput screens that the Broad will perform. The Broad has invested heavily in capital equipment and expertise for chemical biology screens (ie. small molecule drug libraries with robotic delivery and automated screening). These libraries are huge: 50,000-500,000 molecules can be screened. Although much science is hypothesis driven, this kind of large-scale hypothesis-free exploration just hasn’t been possible before. And this certainly isn’t the kind of thing that can be done in a single lab; only dedicated facilities like those at the Broad could carry out this type of “big science.” For collaborators hoping to use the Broad platform, the key appears to be in developing a good automated assay:

Readouts may be anything from classical enzymatic reactions, through FRET for changes in protein interaction, up to subcellular changes captured by automated high-content imaging. An investigator may send a group member to the Broad to take advantage of its resources or may entirely ‘outsource’ assay development to the chaperone. Assay development typically takes two to three months, sometimes up to a year. The assay is then used to screen one or more compound libraries, encompassing at present up to 400,000 substances and growing. (PsychHTS pays for screening a 50,000-compound subset.) ‘Hits’—compounds that affect the assay results in a way that indicates potential usefulness in a psychiatric research context—are automatically retested at several concentrations. The resulting collection of typically between 50 and 500 confirmed hits is then evaluated and prioritized according to criteria of scientific interest and potential drug promise, and thereby winnowed down to the top 10 or 20. The Broad Institute’s organic chemists then synthesize and retest these compounds plus a series of their chemical derivatives, with goals such as improved solubility and more specific binding to putative targets. The goal of the entire procedure is to deliver small-molecule probes that modulate a specific cellular function—essentially tools for subsequent research into the initial hypothesis regarding a psychiatric disease mechanism.

At this point, the new small-molecule probes will need to be tested in animal models of mental illness.

The most appealing aspect is that the Broad is opening up the process to anyone with good ideas for potential screens. The next application deadline is in September. Considering both PsychHTS and the Allen Brain Atlas, is neuroscience moving away from an individual lab model and more toward a “big science” model of projects with lots of collaboration?

A cutting-edge application of the Affy total human exome GeneChip (4X coverage per exon, 40X coverage per gene): Functional and Evolutionary Insights into Human Brain Development through Global Transcriptome Analysis.

From the News and Views, I was intrigued to learn that previous transcriptome analyses of adult human brains found very little difference in gene expression between brain areas:

[...] this suggests that it is the gene expression during development that largely determines higher brain functions by specifying the complexity of neural connections. Numerically, the most important genes relating to cognitive differences between species may be genes that specify how the machinery is put together. In support of this hypothesis, many of the identified differentially expressed genes in this study are related to processes involved in connection formation, such as axonal guidance and cell adhesion.

An impressive 76% of all human genes are expressed in the developing fetal brain. Of those, 33% are differentially expressed over brain regions (13 regions were examined) and 28% are alternatively spliced. The differentially expressed genes are also ones that seem to have evolved the most recently. Even in these early (midgestation) stages, left-right asymmetry was seen, such as the localization of the language-associated FOXP2 genes to Broca’s area.

Of interest to computational folks, they find that gene expression follows power-law scaling (as many other naturally occurring “small-worlds” networks do) with certain hub genes connected to many others and certain spoke genes with relatively few connections. Unsupervised hierarchical clustering is used in this analysis.

A set of two articles recently came out in Science that directly visualize two different (and likely complementary) approaches to synapse specific delivery of gene products. Plasticity at specific synapses (input specificity — we’re restricting the discussion to the dendrites of the post-synaptic neuron) requires proteins (eg. new AMPA receptors) to get to those post-synaptic compartments and membranes. But the intructions for these new proteins are contained in the nucleus with the rest of the genome. Clearly, new proteins synthesized in the soma can’t just be sent everywhere, since only specific inputs (eg. particular dendritic spines) need these new proteins. How does this happen? Hence, the postulated synaptic tag.

Two approaches

Broadly, there are two approaches to synaptic tagging: 1) mRNA is distributed widely and translated locally at tagged locations; 2) protein products are distributed widely in the bodies of dendrites but only contact/off-load at tagged synaptic specializations. This News & Views gives a nice overview of the two papers, which find approach 1) in Aplysia cultures with sensorin mRNA and approach 2) in rat hippocampal neurons with Vesl-1S/Homer-1a protein. It amazes me that both were found pretty much simultaneously, but that might have more to do with the use of the photoconvertible Dendra2 protein than anything else.

With both approaches, we still don’t know why mRNA/protein is directed to a certain location. That is, we can visualize synaptic tagging but we don’t know what is the tag, its ontogeny, or the mechanism of tagging. But that might not be so important to understanding more about neural function. These new tools might allow us to image plasticity at many synapses at once, perhaps even in vivo. But before that, more work is needed… does the optical signal (from the Dendra fusion protein) correlate with degree of potentiation? Can we detect plasticity in the opposite direction, ie. synaptic depression, through another tag?  (As a sidenote to approach 1), the use of 5′ and 3′ UTRs as a sort of molecular zipcode is also intriguing.)

Cryonics never really delivered. But can we now develop the technology to preserve neural structures? Ken Hayworth thinks we can and advocates a brain preservation technology prize. It’s nice to see such big ideas.

I came across this fantastic review of tools for the Genetic Dissection of Neural Circuits in Neuron a few days ago. It’s by Liqun Luo, Ed Callaway, and Karel Svoboda. I highly recommend it, as it spans the gamut from genetic targeting (recombination, binary logic, viral delivery) to circuit reconstruction (super resolution LM, EM, brainbow) to activity modulation and functional mapping (uncaging, artificial GPCRs, light-gated channels, MIST). I don’t think I’ve ever seen quite a review of so many cutting edge neurotechnologies in one place. I can’t recommend this piece enough really. For me, with my lack of molecular expertise, the first sections on combinatorial gene targeting/expression techniques were great, pulling together Gal4, Cre/Flp, and Tet systems into a unified framework, along with more general concepts like site-directed integration, enhancer-trap, and repressor trap (eg. Thy1 mice).

Rewiring the Brain: Inside the New Science of Neuroengineering.

Interviews Boyden and Deisseroth. Follow the link a video of an optogenetically controlled mouse.

Nature methods has a small piece interviewing Karl Deisseroth on properties of the new step ChR2.

Some shortcomings of step ChR2 and future research directions:

Deisseroth expects ongoing efforts to improve key features of these channels. “One disadvantage is that some of the mutants have reduced current compared to wild type, so multiple mutations may help to bring those current levels back up to wild-type levels,” he says. Projects designed to improve membrane targeting and to apply a composite of opsins, including the red light–responsive channelrhodopsin from Volvox carteri, are also in the works in his laboratory.

As has become a hallmark of the Svoboda lab, this new paper in Nature (advance online publication) combines several cutting edge technologies (rAAV-delivered ChR2, most prominently, and 2-photon 1-photon laser stimulation) to do some interesting synaptic physiology.

The subcellular organization of neocortical excitatory connections : Article : Nature.

They used ChR2 (with TTX and 4-AP to block action potentials) to find where on the dendritic tree particular inputs synapsed onto L3 and L5 cells and to measure the strength of those inputs. ChR2 depolarizes the input axon locally (60um spot diameter) at points of (potential) axodendritic contact. If you’ve heard the term “potential synapse” before, then think of this technique as a way of checking potential synapses and seeing if there really is an actual synapse there.

The technique allowed them to map on a L3 barrel cortex pyramidal cell where different thalamic inputs (VPm, POm) and cortical inputs (M1, barrel L2/3, barrel L4):

sCRACM stands for subcellular ChR2-assisted circuit mapping.

PNAS has some interesting articles that I came across today:

1. 3D PALM (open access): Using 2-photon and photoactivatable proteins, the authors image beyond the usual sub-wavelength TIRF limits. They image over multiple microns with 50nm resolution.
2. Neuroeconomics:  Low digit ratio (2d:4d) predicts financial success in traders. Okay, measure the length of your index and ring fingers. (Not sure if this analysis applies for the ladies; the authors only used men in the study.) Calculate the ratio (2d/4d); longer ring fingers signify greater fetal androgen exposure. The mean value is about 0.96. As the authors say,
   

   Digit ratios have been found to predict performance in competitive sports, such as soccer, rugby, basketball, and skiing, so 2D:4D may also predict the risk preferences and physical speed required for high-frequency trading.

   A strong correlation (r~0.5) was found between low digit ratios and profits in short-term trading. So, they take on more risk and make more money. What I want to know is how well the low 2d:4d ratio traders did over the last 6 months!

Nature is reporting on potential flaw in multiple imaging (fMRI) studies of social neuroscience. Ed Vul (a graduate student in my dept) and colleagues have a paper in press that says that many of the high correlations between brain regions and social behavior are implausible, given the inherent variability/noise in fMRI. Furthermore, based on a survey of methods from individual investigators, they created a list of papers that commit, in their view, a statistical mistake (non-independence). Naturally, the authors named in the paper aren’t happy and, according to the Nature article, several rebuttals are in the works. At the very least, to my non-expert eyes, this seems like an important discussion to have about data analysis and methodology.